AI Weather Forecasting Platform Excarta Announces New Models to Expand Capabilities, Provide More Accurate Predictions to Commercial Customers

Building on the momentum from December 2023 launch, Excarta will offer commercial customers in the energy sector and other verticals high-resolution weather forecasts

SAN MATEO Calif., April 24, 2024 /PRNewswire/ -- Today, Excarta announced the release of several new models, allowing the AI-driven weather forecasting platform to provide higher-resolution data-driven forecasts. Compared to traditional models, Excarta has up to 20% lower errors for key weather variables, and the incorporation of the new models will help Excarta offer more detailed forecasts and boost accuracy further. Excarta focuses on commercial use cases, such as load and power forecasting, and extreme event response, but given the models' impressive accuracy, they also have wider applications. 

"It's exciting to see how quickly AI weather models are evolving," said Vivek Ramavajjala, CEO and Founder of Excarta. "Our latest models combine the best of graph-based models and vision transformers, resulting in memory- and compute- efficient models that outperform even the best conventional forecasts. We're also starting to use more real-time data to further improve the models. Incorporating new modes of data allows us to produce forecasts more relevant for specific applications and use-cases." 

Excarta consistently outperforms traditional weather models and captures signals missed by other forecasts. For instance, conventional models typically offer hourly forecasts only for the first three to five days. Excarta's models are able to provide hourly forecasts for fourteen days, including for variables not predicted by other AI weather models, such as solar irradiance, wind speeds at hub height, precipitation amounts and types, etc. Even compared to high-resolution regional models, which are tuned for specific geographies, Excarta is able to provide as much as a 15% lower error rate on key variables. 

In addition, Excarta harnesses the low cost of AI models to run ensemble forecasts, providing customers with an estimate of uncertainty in the forecast. This allows customers to incorporate weather-induced uncertainty in their planning . While impressive on their own, Excarta's AI models powerfully complement traditional models ? combining Excarta's AI forecasts with data from traditional models showed an impressive 25% lower error rate compared to traditional models alone.

Excarta's models can be used to replace or complement existing forecasting methods, making for an easy transition to leveraging AI-powered accurate weather information. And accuracy is becoming increasingly important as the impacts of climate change become more apparent. 

"As climate change makes weather more volatile, AI weather forecasting will be key to climate resiliency efforts," added Ramavajjala. "Whether it's evacuation routes for severe weather events, identifying where to allocate resources in advance of storms, or boosting solar and wind energy production, weather forecasts help determine our actions. So we want to make sure that decision-makers have the most accurate information possible."

Excarta is currently available for commercial customers. To learn more, visit excarta.io. 

About Excarta

Founded in 2022, Excarta was formed to commercialize the power of AI for business-targeted weather forecasting. Excarta enables businesses to easily, quickly and inexpensively produce highly accurate, high-resolution weather forecasts for the parameters they care about to gain greater efficiencies and resiliency while mitigating disasters. Excarta offers a cloud-based platform that can utilize unlimited amounts of weather data to study and forecast weather for geographies a fraction of traditional size with great accuracy and covering greater time spans. Learn more about Excarta or request a demo through the company website.

SOURCE Excarta, Inc